yaojingang/GEOInflow

What it solves

GEOInflow helps website owners understand how AI crawlers (like those from OpenAI, Google, or Perplexity) interact with their content. It bridges the gap between standard web traffic analytics and the "black box" of AI search and training, allowing teams to see which pages are being crawled, whether those crawls are successful, and how AI-driven referrals impact their business.

How it works

The system collects data from three primary sources: browser scripts on your website, server or CDN access logs, and your own content inventories (like sitemaps). It processes this data to identify specific crawler identities, track request patterns, and monitor for "blind spots" where important content might be missed. All data is stored and analyzed within your own self-hosted infrastructure using PostgreSQL and Redis.

Who it’s for

It is designed for media publishers, content-heavy websites, developer documentation teams, and GEO (Generative Engine Optimization) professionals who need to manage and monitor multiple sites. It provides tools for both operators to view traffic reports and maintainers to inspect data ingestion and system health.

Highlights

• Tracks AI crawler activity, including identity, purpose, and request volume.
• Compares observed crawl data against your own content inventories to identify coverage gaps.
• Monitors crawl success rates, latency, and potential access errors.
• Integrates standard website analytics like pageviews, visitor paths, and conversion tracking.
• Supports self-hosting via Docker with built-in data privacy controls.
• Provides clear evidence-based reporting to distinguish between training, search, and user-triggered requests.